Cross layer optimization of wireless sensor and sensor-actuator networks

Télécommunicationsde Paris

THÈSE

Présentée pour obtenir le grade de docteur

de l'École Nationale Supérieure des Télécommunications

Spécialité : Informatique et Réseaux

Presente par:

Muhammad Farukh MUNIR

Cross-Layer Optimizations of Wireless Sensor

and Sensor-Actuator Networks

Souténu publiquement le26 Fevrier 2009 devant lejury composéde :

Président : Elie NAJM Telecom ParisTech,France

Rapporteurs: Michel DIAZ CNRS,France

Congduc PHAM Universitede Pau,France

Examinateurs : Isabelle GUÉRIN-LASSOUS Universitede Lyon 1

Mischa DOHLER CTTC, Barcelona,Spain

Directeur de thèse : Fethi FILALI EURECOM, France

(14:30 -EURECOMSophia-Antipolis)

capteurs et capteurs-actionneurs sans-l

Muhammad Farukh MUNIR

Cross-Layer Optimizations of Wireless Sensor and

Sensor-Actuator Networks

À ma femme etma famille...

Je remercie mon directeur de these Fethi Filali pour avoir accepte mon travail, pour son

aide precieuse, technique et morale, et pour ca grande patience durant toutes les phase de

cettethese.

Je remercie HEC(Higher Education Comission,Pakistan) pour leur supportnanciere.

Je remercie egalement tousles doctorants et tousles personnel de EURECOMpour leur

sympathieetlabonne ambiance qu'ilsgenerent ausein de l'institut.

Enn, ce travail n'aurait pas pu etre accomplie son l'amour et le soutien de toute ma

famille, mafemme,mes parents, mes soeurs, etmonfrere.

MuhammadFarukhMUNIR

Sophia-Antipolis

26 February2009

1 Introduction 15

1.1 General Introduction . . . 15

1.2 Applications . . . 20

1.2.1 Military Applications. . . 20

1.2.2 Civil Applications . . . 21

1.2.3 EnvironmentalApplications . . . 21

1.2.4 Medical Applications . . . 22

1.3 Motivations and Objectives . . . 22

1.4 Thesis Outline and Contributions . . . 25

2 Cross-Layer Routing in WSNs 29 2.1 Introduction . . . 29

2.2 Related Literature . . . 33

2.3 Problem Statement . . . 36

2.4 Data Collection Mechanism . . . 36

2.4.1 OpenSystem (LayeredArchitecture) . . . 37

2.4.2 Closed System (Cross-LayeredArchitecture) withSingle Transmit Queue 37 2.4.3 Applications for ClosedSystem withSingle Transmit Queue . . . 38

2.4.4 Closed System withtwoTransmitQueues . . . 38

2.4.5 An Example . . . 39

2.5 StabilityAnalysis . . . 43

2.5.1 OpenSystem . . . 43

2.5.2 Closed System withSingleTransmit Queue . . . 46

2.5.3 Closed System withtwoTransmitQueues . . . 48

2.6 Routing Algorithmsfor Dierent SystemsUnderConsideration . . . 51

2.6.1 OpenSystem . . . 51

2.6.2 Closed System withSingleTransmit Queue . . . 52

2.6.3 Practical Considerations . . . 52

2.7 SimulationResults . . . 52

2.7.1 OpenSystem Stability . . . 54

2.7.2 Closed System Stability . . . 55

2.7.3 OpenSystem Routing . . . 56

2.7.4 Closed System Routing . . . 57

2.7.5 Closed System withTwo Transmit Queues . . . 59

2.8 Conclusions and FutureWork . . . 60

3 Cross-layer Routing in SANETs 69

3.1 Introduction . . . 69

3.2 Related Literature . . . 71

3.3 Problem Statement . . . 72

3.4 The Network Model . . . 73

3.4.1 Channel Modeland Antennas . . . 73

3.4.2 Frequency . . . 73

3.4.3 Neighborhood RelationModel. . . 73

3.4.4 Application-layer Sampling-Mechanism. . . 74

3.4.5 Relaying . . . 74

3.4.6 Trac Model . . . 75

3.4.7 Channel AccessMechanism . . . 75

3.5 Optimization Problemfor Open System . . . 75

3.5.1 Lagrange Dual Approach . . . 79

3.5.2 Deterministic Primal-Dual Algorithm. . . 79

3.6 StochasticDelayControl And StabilityUnderNoisyConditions . . . 80

3.6.1 StochasticPrimal-Dual Algorithm For Delay Control . . . 80

3.6.2 Probability OneConvergence OfStochasticDelayControl Algorithm . . 80

3.7 Rateof Convergence of Stochastic Delay Control Algorithm . . . 82

3.8 Sensor-Actuator Coordination . . . 83

3.8.1 Optimal Actuator Selection . . . 83

3.8.2 A Distributed RoutingAlgorithm . . . 85

3.9 Actuator-Actuator Coordination . . . 85

3.9.1 Classication of Actuation Process . . . 86

3.9.2 Data Collection andDistributed Routing . . . 86

3.9.3 StabilityAnalysis withPowerControl . . . 87

3.9.4 Dynamic Actuator Cooperation . . . 88

3.10 Implementation Results . . . 89

3.10.1 Optimization inOpen System . . . 91

3.11 Conclusions and FutureWork . . . 93

4 The LEAD Cross-Layer Architecture for SANETs 97 4.1 Introduction . . . 98

4.2 Related Literature . . . 101

4.3 Problem Statement . . . 105

4.4 NetworkModel . . . 106

4.4.1 Channel Model . . . 106

4.4.2 Neighborhood RelationModel. . . 106

4.4.3 Forwarding (Relaying) . . . 106

4.4.4 Channel Modeland Antennas . . . 107

4.4.5 Frequencyand MAC . . . 108

4.5 The Three-Level Coordination Framework For SANETs . . . 108

4.6 LEAD-RP: TheLEADRouting Protocol . . . 110

4.6.1 PowerConsumptionModel . . . 110

4.6.2 Actuator-Selection andOptimal ow Routing . . . 110

4.7 LEAD-ADP: The LEADActuator Discovery Protocol . . . 116

4.7.1 The Learning-phase . . . 116

4.7.2 The Coordination-phase . . . 119

4.7.3 FailureandRecovery-phase . . . 119

4.8 Deterministic Lifetime Maximization . . . 120

4.8.1 Lagrange Dual Approach . . . 120

4.8.2 Deterministic Primal-Dual Algorithm. . . 122

4.9 LEAD-MAC: TheLEAD MediumAccessControl . . . 123

4.9.1 NetworkLearning Phase . . . 124

4.9.2 Scheduling Phase . . . 124

4.9.3 Adjustment Phase . . . 126

4.10 LEAD-Wakeup . . . 126

4.10.1 Adaptivityto NetworkConditions . . . 126

4.10.2 Analysis ofLEADWakeup . . . 126

4.11 Actuator to SensorTransmission Schemes . . . 128

4.11.1 Transmission at asingle frequency(Reuse Factor 1) . . . 128

4.11.2 Transmissionsat dierent frequencies (HigherReuseFactor) . . . 129

4.11.3 Actuator Cooperation(Joint Beamforming) . . . 129

4.12 SimulationResults . . . 130

4.13 Conclusions and Futurework . . . 139

5 Cross-Layer Routing in UASNs 143 5.1 Introduction . . . 143

5.2 Related Work . . . 144

5.3 The DesignCriteria . . . 145

5.3.1 Gold Sequences . . . 145

5.3.2 The Timereversal(phase conjugation)approach . . . 145

5.3.3 Underwater PropagationModel . . . 146

5.4 Case I: Single-Hop Communication Framework . . . 147

5.4.1 Waveform design . . . 148

5.4.2 Pulse positionmodulation (PPC-PPM) . . . 148

5.4.3 Calculation of SNRand BER . . . 149

5.4.4 SimulationResults . . . 151

5.5 Case II:Multi-Hop Communication Framework . . . 156

5.5.1 A Three-Node LinearNetwork . . . 156

5.5.2 NetworkModel . . . 158

5.5.3 The RoutingAlgorithm . . . 158

5.5.4 SimulationResults . . . 159

5.6 Conclusions and FutureWork . . . 161

6 Conclusion and Outlook 163 6.1 Summary of Contributions . . . 163

6.2 Future Directions . . . 166

7 Résumé en Francais 167

1.1 A Wireless SensorNetwork . . . 16

1.2 A Wireless Sensor-Actuator Network . . . 17

1.3 A Layeredand Cross-Layered Architecture . . . 25

1.4 A System withTwo-Queues at MAC . . . 26

1.5 The LEADFramework . . . 27

2.1 FlowSplitting . . . 31

2.2 Medium AccessControl . . . 32

2.3 NetworkConguration . . . 40

2.4 An example Network Topology . . . 43

2.5 Markovchainfor the expectednumberof packetsat node

i

^{,case 1:}

P l φ _l,i = 0

^{. 46} 2.6 Markovchainfor theexpectednumberof packetsat node

i

^,case2:

P l φ _l,i > 0

^{. 47} 2.7 NetworkSimulated for Stability . . . 53

2.8 Sensor network architecture.

→

^{representstheowof packets fromthesource} tothedestination. Theforwardingsensornetworkreceivesapacketandqueues into the forwarding queueattheMAClayer. Therouting layerdoesnotbuer theforwarding trac. . . 54

2.9 Delays incurred on routes

2 → 5 → 0

^,

2 → 1 → 0

^{for Open System. Where}

λ 1 = λ 2 = λ 3 = λ 4 = λ 5 = 0.2

^{. . . . . . . . . . . . . . . . . . . . . . . . . . . 55}

2.10 Delays incurred on routes

4 → 3 → 0

^,

4 → 1 → 0

^{for Open System. Where}

λ ₁ = λ ₂ = λ ₃ = λ ₄ = λ ₅ = 0.2

^{. . . . . . . . . . . . . . . . . . . . . . . . . . . . 56}

2.11 Delays incurred on routes

2 → 5 → 0

^,

2 → 1 → 0

^{for Closed System. Where}

λ ₁ = λ ₂ = λ ₃ = λ ₄ = λ ₅ = 0.2

^{. . . . . . . . . . . . . . . . . . . . . . . . . . . 57}

2.12 Delays incurred on routes

4 → 3 → 0

^,

4 → 1 → 0

^{for Closed System. Where}

λ ₁ = λ ₂ = λ ₃ = λ ₄ = λ ₅ = 0.2

^{. . . . . . . . . . . . . . . . . . . . . . . . . . . . 58}

2.13 NetworkSimulated for Routing . . . 59

2.14 Delays incurred on routes

3 → 1 → 0

^,

3 → 2 → 0

^,

5 → 1 → 0

^,

5 → 4 → 0

for open system.

α 1 = 0.2, α 2 = 0.15, α 3 = 0.1, α 4 = 0.2, α 5 = 0.2

^,

λ 1 = 0.01, λ ₂ = 0.01, λ ₃ = 0.04, λ ₄ = 0.05, λ ₅ = 0.05.

^{. . . . . . . . . . . . . . . . . . 60}

2.15 Delays incurred on routes

3 → 1 → 0

^,

3 → 2 → 0

^,

5 → 1 → 0

^,

5 → 4 → 0

for open system.

α ₁ = 0.1, α ₂ = 0.1, α ₃ = 0.1, α ₄ = 0.1, α ₅ = 0.1

^,

λ ₁ = 0.01, λ ₂ = 0.05, λ ₃ = 0.05, λ ₄ = 0.01, λ ₅ = 0.04.

^{. . . . . . . . . . . . . . . . . . 61}

2.16 Trac splitover theroutes

3 → 1 → 0

^,

3 → 2 → 0

^,

5 → 1 → 0

^,

5 → 4 → 0

^for open system. . . 62

2.17 Delays incurred on routes

3 → 1 → 0

^,

3 → 2 → 0

^{for closed system with}

λ ₁ = 0.1, λ ₂ = 0.2, λ ₃ = 0.1, λ ₄ = 0.005, λ ₅ = 0.1.

^{. . . . . . . . . . . . . . . . 62}

2.18 Delays incurred on routes

5 → 1 → 0

^,

5 → 4 → 0

^{for closed system with}

λ ₁ = 0.1, λ ₂ = 0.2, λ ₃ = 0.1, λ ₄ = 0.005, λ ₅ = 0.1.

^{. . . . . . . . . . . . . . . . 63}

2.19 Trac splitover the routes

3 → 1 → 0

^,

3 → 2 → 0

^,

5 → 1 → 0

^,

5 → 4 → 0

^for closed system. . . 63

2.20 Convergence ofchannel access rates for closed system. . . 64

2.21 Expected Delayinarandomly deployed networkovertime . . . 65

2.22 CDF ofthe Estimated Delayina randomlydeployed network . . . 66

2.23 AverageDelays forTwo-Queues Vs. Single QueueSystem . . . 67

3.1 Architecture of Sensor-Actuator Networks . . . 74

3.2 The Simulated Networkconsistingof 7sensors and 2actuators. . . 90

3.3 Throughput vs. Actuator Density . . . 92

3.4 Energy Consumption forControl Overhead . . . 92

3.5 A SimpleNetworkTopology . . . 93

3.6 Convergence of

µ ₃

^usingdistributed primal-dualalgorithm . . . 94

3.7 Convergence of

µ ₄

^usingdistributed primal-dualalgorithm . . . 95

3.8 Convergence of

µ 5

^usingdistributed primal-dualalgorithm . . . 96

3.9 Convergence of

µ ₆

^usingdistributed primal-dualalgorithm . . . 96

4.1 Architecture of SANETs . . . 107

4.2 The LEADArchitecture . . . 109

4.3 AttachRequest bysensors at the start ofADP . . . 117

4.4 Actuator-replies (AttachReply) forcorrespondingAttachRequest messages . . . 119

4.5 The Local Cluster formulated at thetermination ofADP. . . 120

4.6 A Self-Organized Tree(SOT) . . . 122

4.7 The occurrence offailureand steps required for recovery procedure. . . 122

4.8 Energy Savingsthroughadaptive dutycycle . . . 127

4.9 Networklifetime underanalyticaland simulation results . . . 131

4.10 Mean end-to-endtransmission delays . . . 132

4.11 Mean energy consumption asafunction of timefor anetwork of 100 sensors . . 133

4.12 Mean numberof transmissionsperend-to-endpath (mean path length) . . . . 134

4.15 AverageNumberof isolatedSensors vs. Transmit Power. . . 134

4.13 Averagedelayinacluster

→

^{increasing #ofnodes . . . . . . . . . . . . . . . . 135}

4.14 Averageenergy consumption ina cluster

→

^{increasing #of nodes . . . . . . . . 136}

4.16 AverageNumberof isolatedSensors Vs. Total Numberof DeployedSensors. . . 136

4.17 Probability ofSensor Inactivityintheareas of thesensingeld for thecaseof Reuse Factor 1ScheduleBroadcast Transmission. . . 137

4.18 Probability ofSensor Inactivityintheareas of thesensingeld for thecaseof Reuse Factor 3ScheduleBroadcast Transmission. . . 138

4.19 Probability ofSensor Inactivityintheareas of thesensingeld for thecaseof joint Maximal Ratio Combining Beamforming. . . 138

5.1 Passive Phase Conjugation(PPC) . . . 146

5.2 Waveform Designfor PPC-PPM . . . 149

5.3 BlockDiagramof PPC-PPM usingGold sequences . . . 152

5.4 Bit-Error-Rate Vs. SNRfor 126 [bps] . . . 152

5.5 Bit-Error-Rate Vs. SNRfor 500 [bps] . . . 153

5.6 Bit-Error-Rate Vs. Distance (m) . . . 154

5.7 Bit-Error-Rate Vs. Physical layerRate . . . 155

5.8 Bit-Error-Rate Vs. Depth . . . 156

5.9 Passive Phase Conjugation(PPC) ina 3-node network . . . 157

5.10 Packet-Error-Rate Vs. SNR . . . 160

5.11 Averagenumberof packettransmissionattempts . . . 161

7.1 Un réseau descapteurssans-l . . . 168

7.2 Un réseau sanslcapteurs-actionneurs . . . 170

7.3 Une architecture en couchesetinter-couches . . . 183

7.4 Un systèmeavec deuxles d'attente au-MAC . . . 184

7.5 The LEADFramework . . . 188

2.1 Node levelDelays . . . 42

2.2 FlowlevelDelays . . . 42

2.3 Results onThroughput andStabilityRegion . . . 64

3.1 Comparisonbetweentheresultsoftheproposedprimal-dual algorithmandthe

theoretical optimalsolution . . . 93

4.1 Notations . . . 140

4.2 Useful statesfor thesensornodewithassociatedpower consumptionand delay

(time to reach

S ₄

^{fromanygiven state) . . . . . . . . . . . . . . . . . . . . . . 141}

4.3 The simulation area is such that there are atleast two sensors in each others

transmission range . . . 141

5.1 Parameters . . . 147

Introduction

1.1 General Introduction

Recentadvancesinmicro-electro-mechanical systems(MEMS) technology,wirelesscommuni-

cations, and digital electronics have enabled thedevelopment of low-cost, low-power, multi-

functionalsensornodesthataresmallinsize andcommunicate untetheredinshortdistances.

Wireless Sensor Networks (WSNs) consist of large number of distributed sensor nodes that

organizethemselves into amultihop wireless network asshown inFigure1.1. Each nodehas

one or more sensors, embedded processors, and low-power radios, and is normally battery

operated. Typically, these nodes coordinate to perform a common task. These tiny sensor

nodes, which consist of sensing, data processing, and communicating components, leverage

the idea of WSNs based on collaborative eort of a large number of nodes. WSNs represent

a signicant improvement over traditional sensors, which are deployed in the following two

ways [1]:

•

^{Sensors can be positioned far from the actual} phenomenon, i.e., something known by sense perception. In this approach, large sensors that use some complex techniques to

distinguish the targetsfrom environmental noise arerequired.

•

^{Severalsensorsthatperformonly sensingcanbedeployed. Thepositions ofthesensors}

and communications topology can be carefully engineered. They transmit time series

of thesensed phenomenon to thecentral nodeswhere computations areperformed and

data are fused. The central entity is shown as sink in Figure 1.1. It can be placed

anywhere depending upon the application needs.

Asensornetworkiscomposedofalargenumberofsensornodes,whicharedenselydeployed

either inside the phenomenon or very close to it. The position of sensor nodesneed not be

engineered or pre-determined. This allows random deployment in inaccessible terrains or

disaster relief operations. Onthe other hand, this also means thatsensor network protocols

and algorithms must possess self-organizing capabilities. Another unique feature of WSNs is

the cooperative eort of sensor nodes. Sensor nodes are tted with an on-board processor.

Insteadofsending therawdatato thenodesresponsibleforthefusion,sensornodesusetheir

processing abilities to locally carryout simple computations and transmit only the required

and partially processeddata.

The above described features ensure a wide range of applications for WSNs. Some of

the application areasare health, military, environment, civil, and security. For example,the

Figure1.1: A Wireless SensorNetwork

physiologicaldataaboutapatientcanbemonitoredremotelybyadoctor. Whilethisismore

convenientfor thepatient,italsoallows thedoctortobetterunderstandthepatient'scurrent

condition. WSNs can alsobeusedto detect foreignchemical agents intheair andthewater.

Theycan helpidentifythetype,concentration, andlocation of pollutants. Inessence, WSNs

can provide the end user with intelligence and a better understanding of the environment.

We envision that, in future, WSNs will be an integral part of our lives, more so than the

present-day personal computers.

Theselowpowerandlossynetworks(LLNs) aremadeupof manyembedded deviceswith

limited power, memory, and processing resources. They are interconnected by a variety of

links, such as IEEE 802.15.4, Bluetooth, Low Power WiFi, wired or other low power PLC

(PowerlineCommunication)links. LLNsaretransitioning toan end-to-endIP-basedsolution

to avoid the problem of non-interoperable networks interconnected by protocol translation

gateways and proxies. Existing routing protocols such as OSPF, IS-IS, AODV, and OLSR

have been evaluated by the IETF ROLL [2 ] working group and have in their current form

beenfoundtonot satisfyallofthespecicWSNroutingrequirements. Thegroupiscurrently

workingon the standardization ofrouting funtionalityfor thespecic requirements posedby

LLNs.

Wireless sensor-actuator networks 1

(SANETs) are among the most addressed research

eldsintheareaofinformationandcommunicationtechnologies(ICT)thesedays,intheUS,

EuropeandAsia. SANETsarecomposedofpossiblyalargenumberoftiny,autonomoussensor

devicesandactuators 2

equipped withwireless communicationcapabilities asshown inFigure

1.2. Oneofthemostrelevantaspectsofthisresearcheldstandsinitsmultidisciplinarityand

thebroadrangeof skillsthatareneededto approach their design. Theoryof control systems

isinvolved, networking, middleware, application layer issues arerelevant, joint consideration

of hardware and software aspects is needed, and their use can range from biomedical to

industrial or automotive applications, from military to civil environments, etc. Distributed

1

In related literature, the termWSANs (Wireless Sensor-Actor Networks) is also used to represent the

same.

2

Inrelevantliterature,theterm'actor' isusedto representthesame,i.e.,adevicethathasbothcommu-

nicationandactuationcapabilities.

systems based on networked sensors and actuators with embedded computation capabilities

enable an instrumentation of thephysical world at an unprecedented scale and density, thus

enabling a new generationof monitoring and control applications. SANETs arean emerging

technologythathasawide rangeofpotential applicationsincludingenvironment monitoring,

medical systems,robotic exploration, and smart spaces. SANETs are becomingincreasingly

importantinrecentyearsduetotheirabilitytodetectandconveyreal-time,in-situinformation

for many civilianand military applications.

Each sensor node has one or more sensors (including multimedia, e.g., video and audio,

or scalar data, e.g., temperature, pressure, light, infrared, and magnetometer), embedded

processors, low-power radios, and is normally battery operated. An actuator is a device to

convert an electrical control signal to a physical action, and constitutes the mechanism by

which an agent acts upon thephysical environment. From theperspective consideredin this

thesis, however, an actuator, besides being able to act on the environment by means of one

or severalactuators, isalsoanetwork entitythatperformsnetworking-relatedfunctionalities,

i.e., receive, transmit, process, and relay data. For example, a robot may interact with the

physical environment by means of several motors and servo-mechanisms (actuators). How-

ever, from a networking perspective, the robot constitutes a single entity, which is referred

to as actuator. Hence, the term actuator embraces heterogeneous devices including robots,

unmanned aerialvehicles (UAVs),and networked actuators suchaswater sprinklers,pan/tilt

cameras, robotic arms,etc. Applicationsof SANETsmayinclude team of mobilerobotsthat

perceive theenvironment from multiple disparate viewpoints based on the datagathered by

a sensor network, a smart parking system that redirects drivers to available parking spots,

or a distributed heating, ventilating, and air conditioning (HVAC) system basedon wireless

sensors.

Figure1.2: A WirelessSensor-Actuator Network

However, due to the presenceof actuators, SANETs have some dierencesfrom WSNsas

outlined below:

•

^{Whilesensornodesaresmall,}inexpensivedeviceswithlimitedsensing,computationand wirelesscommunicationcapabilities,actuatorsareusuallyresource-richdevicesequipped

withbetterprocessingcapabilities,strongertransmissionpowersandlongerbatterylife.

•

^{In SANETs, depending on the} application there may be a need to rapidly respond to sensor input. Moreover, to provide right actions, sensor data must still be valid at

thetimeofacting. Therefore,theissueofreal-time communicationisveryimportantin

SANETssinceactionsareperformedontheenvironment aftersensingoccurs. Examples

can be are application where actions shouldbe initiated onthe event area assoonas

possible.

•

^{Thenumberofsensornodesdeployed instudyingaphenomenonmaybeintheorderof}

hundredsorthousands. However,suchadensedeployment isnot necessaryforactuator

nodes due to the dierent coverage requirements and physical interaction methods of

acting task. Hence,inSANETs thenumberofactuatorsismuchlowerthan thenumber

of sensors.

•

^{In order to provide eective sensingand acting, a} distributedlocal coordination mech- anism is necessary among sensors and actuators. In WSNs, the central entity (i.e.,

sink) performsthefunctionsofdata collectionand coordination. Whereas,inSANETs,

newnetworkingphenomenacalledsensor-sensor,sensor-actuator,andactuator-actuator

coordination may occur. In particular, sensor-sensor coordination deals with local col-

laboration amongneighbors toperform in-network aggregation and exploit correlations

(both spatialand temporal). Sensor-actuatorcoordination providesthetransmissionof

event features from sensors to actuators. After receiving event information, actuators

mayneedtocoordinate(actuator-actuator coordination)witheachother(dependonthe

acting application) inorder to make decisions onthe most appropriate way to perform

theactions.

Many protocols and algorithms have been proposed for WSNs in recent years [3]. However,

since the above listed requirements impose stricter constraints, they may not be well-suited

for theinherent featuresand application requirements ofSANETs. Moreover, although there

has been some research eort related to SANETs,to the best ofour knowledge, almost none

oftheexistingstudiestodateinvestigateresearchchallengesoccurringduetotheco-existence

of sensorsand actuators.

Oceanbottomsensornodesaredeemed toenableapplicationsforoceanographicdatacol-

lection,pollutionmonitoring,oshoreexploration,disasterprevention,assistednavigationand

tactical surveillance applications. Multiple Unmanned or Autonomous Underwater Vehicles

(UUVs, AUVs), equipped with underwater sensors, will also nd application in exploration

of natural undersea resources and gathering of scientic data in collaborative monitoring

missions. To make these applications viable, there is a need to enable underwater commu-

nications among underwater devices. Underwater sensor nodes and vehicles must possess

self-conguration capabilities, i.e., they must be able to coordinate their operation by ex-

changing conguration, location and movement information, and to relay monitored data to

an onshore station.

Wirelessunderwateracousticnetworkingistheenabling technologyfortheseapplications.

Underwater Acoustic Sensor Networks (UASN) consist of a variable number of sensors and

vehicles that are deployed to perform collaborative monitoring tasks over a given area. To

achievethis objective,sensorsandvehiclesself-organize inanautonomousnetwork which can

adapt to thecharacteristics ofthe ocean environment [5].

Underwaternetworking isarather unexplored areaalthoughunderwater communications

have been experimented since World War II, when, in 1945, an underwater telephone was

developed inthe United States to communicate withsubmarines. Acoustic communications

are the typical physical layertechnology in underwater networks. In fact, radio waves prop-

agate at long distances through conductive sea water only at extra low frequencies (30-300

Hz), which require large antennae and high transmissionpower. Optical waves do not suer

from such highattenuation but are aectedbyscattering. Moreover, transmission of optical

signals requireshighprecision inpointingthenarrow laserbeams. Thus, linksinunderwater

networksarebasedon acousticwireless communications.

The traditional approach for ocean-bottom or oceancolumn monitoring is to deployun-

derwatersensorsthatrecorddataduringthemonitoringmission,andthenrecovertheinstru-

ments. This approach hasthefollowing disadvantages:

•

^{Realtimemonitoring isnotpossible. Thisiscriticalespeciallyin}surveillanceorinenvi- ronmentalmonitoringapplicationssuchasseismicmonitoring. Therecordeddatacannot

beaccesseduntilthe instrumentsarerecovered,whichmayhappenseveralmonthsafter

thebeginning ofthe monitoring mission.

•

^No interaction is possible between onshore control systems and the monitoring instru- ments. This impedes any adaptive tuning of the instruments, nor is it possible to

recongure the systemafterparticular eventsoccur.

•

^{If failures or} miscongurations occur, itmay not be possible to detect them beforethe instruments arerecovered. Thiscan easily leadto thecomplete failure of amonitoring

mission.

•

^{Theamountofdatathatcanberecordedduringthemonitoringmissionbyeverysensor}

is limited bythe capacity ofthe onboard storagedevices(memories, hard disks,etc.).

Therefore,thereisaneedtodeployunderwaternetworksthatwillenablerealtimemonitoring

ofselected oceanareas, remoteconguration and interaction withonshore human operators.

Thiscan beobtained byconnectingunderwater instruments bymeans ofwireless linksbased

on acoustic communication.

Many researchers arecurrently engaged indeveloping networking solutions for terrestrial

WSNs. AlthoughthereexistmanyrecentlydevelopednetworkprotocolsforWSNs,theunique

characteristics oftheunderwateracousticcommunicationchannel, suchaslimitedbandwidth

capacity and variable delays, require for very ecient and reliable new data communication

protocols. Thequalityoftheunderwateracousticlinkishighlyunpredictable,since itmainly

depends on fading and multipath, which are not easily modeled phenomena. This, in re-

turn, severely degrades theperformance at higher layers such asextremely long andvariable

propagation delays. In addition, this variation is generally larger inhorizontal links than in

vertical ones. Acoustic signaling for wireless digital communications in the sea environment

can be a very attractive alternative to both radio telemetry and cabled systems. However,

time-varyingmultipathandoftenharshambient noiseconditionscharacterize theunderwater

acoustic channel, oftenmakingacousticcommunicationschallenging. Major challenges inthe

design of UASNsare:

•

^{The channel isseverelyimpaired,mainly due to multipath.}

•

^{Temporary lossof} connectivity mainly dueto shadowing.

•

^{Thepropagation delay isve ordersofmagnitude higherthaninradiofrequencyterres-}

trial channels andis usuallyvariable[4 ].

•

^{Extremely low availablebandwidth.}

•

^{Limited battery energy at disposal.}

Sinceunderwatermonitoringmissionscanbeextremelyexpensiveduetothehighcostinvolved

in underwater devices, itis important thatthe deployed network be highlyreliable, soas to

avoid failureofmonitoring missions dueto failureof singleor multiple devices. For example,

it is crucial to avoid designing the network topology with single points of failure that could

compromise the overall functioning of the network. The network capacity is also inuenced

by the network topology. Since the capacity of the underwater channel is severely limited,

it is very important to organize the network topology such a way that no communication

bottlenecks areintroduced.

1.2 Applications

TherangeofapplicationsofWSNs,SANETs,andUASNsareincreasingveryfastandcovering

several domains: military,civil,environmental, health, etc. In this section, we will talkmore

about such applicationsineach ofthese domains[6].

1.2.1 Military Applications

•

^Asset Monitoring: commanders can monitor locationsof thetroops, weaponsand sup- plies toenhance thecontroland communication.

•

^BattleeldMonitoring: vibrationandmagneticsensorscanlocateandtrackenemyforces inthebattleeld.

•

^{Urban Warfare: deploying sensors in cleared buildings can prevent their} reoccupation and track theenemyactivityinside them.

•

Protection: prevention andprotectionfromradiations,biologicalandchemicalweapons can be achieved by the deployment of a WSN, which detects the level of radiation or

thepresence oftoxic products.

•

Distributed Tactical Surveillance: AUVs and xed underwater sensors can collabora- tively monitor areas for surveillance, reconnaissance, targeting and intrusion detection

systems. For example, in[7],a 3Dunderwatersensornetwork is designed for a tactical

surveillancesystemthatisabletodetectandclassifysubmarines,smalldeliveryvehicles

(SDVs) and divers based on thesensed data frommechanical, radiation, magnetic and

acoustic microsensors. With respect to traditional radar/sonar systems, UASNs can

reachahigher accuracy,and enabledetectionandclassication of lowsignaturetargets

byalso combining measures fromdierent typesof sensors.

•

^Mine Reconnaissance: Thesimultaneous operation of multiple AUVs withacoustic and opticalsensorscan beusedtoperformrapidenvironmentalassessmentanddetectmine-

like objects.

1.2.2 Civil Applications

•

Surveillance: a sensor network can detect re in buildings and give information about its location. It can also detectintrusions and trackhumanactivity.

•

^Disaster Prevention: sensor nodes deployed under water can prevent fromdisaster like oceanic earthquakeor impendingtsunami.

•

^{Smart Metering Solutions: smart metering solutions, provided by coronis, based on}

wavenis [8] wireless technology have been deployed in millions of residential, industrial

and commercial installations aroundtheworld, linking consumers'gas, water andelec-

tricitymetersecientlywithoperator'sback-endinformationandbillingsystems. These

advancedsolutionsareusedforwirelesswalk-by,drive-byandfullyautomatedxednet-

workmetering. Waveniswirelesstechnologyprovidestheultra-longrangeandextremely

lowpowerconsumptionthatareessential foreectivelast-mile,outdoorcoverageinme-

tering networksthatserve entire cities,includingdenseurbanareasaswellassprawling

suburban andcommercial zones.

•

^Assisted Navigation: sensors can be used to identify hazards on the seabed, locate dangerous rocks orshoalsinshallowwaters, mooringpositions,submerged wrecks,and

to perform bathymetryproling.

•

^{Disaster Recovery: after an earthquake or a terrorist attack, sensor nodes can detect}

signsof life insideadamaged building.

•

^{Smart Park: a} distributed control system supported by SANET. It improves mobility intheurbanareabyndingfreeparkingspotsfordriverswillingto park[9 ,10]. Italso

decreases the risk ofpossible accidents, pollution, andeliminate roadrage.

1.2.3 Environmental Applications

•

Environment and Habitat Monitoring: a WSN deployed in a sub-glacial environment [11 , 12 ]cancollectinformation about icecapsand glaciers. WSNscan alsobedeployed

to measure population of birds and other species [13 ]. Also,WSN can provide a ood

warning[14 ] and monitorcoastalerosion [15 ].

•

^{DisasterDetection: forestrecanbedetectedandlocalizedbyadenselydeployedWSN.}

•

^{Ocean Sampling Networks: networks of sensors and AUVs, such as the} Odyssey-class AUVs[16 ],canperformsynoptic,cooperativeadaptivesamplingofthe3Dcoastalocean

environment [17]. ExperimentssuchastheMontereyBay eldexperiment [18 ] demon-

strated the advantages of bringingtogether sophisticated new roboticvehicleswithad-

vancedoceanmodelsto improvetheabilitytoobserve andpredictthecharacteristicsof

theoceanic environment.

•

Environmental Monitoring: UASNs can perform pollution monitoring (chemical, bio- logical and nuclear). For example, it may be possible to detail the chemical slurry of

antibiotics, estrogen-type hormones and insecticides to monitor streams, rivers, lakes

and ocean bays (water quality in-situ analysis) [19]. Monitoring of oceancurrents and

winds, improvedweatherforecast,detectingclimatechange,understandingandpredict-

ing the eect of human activities on marine ecosystems, biological monitoring such as

tracking of shes or micro-organisms, are other possible applications. For example, in

[20 ], the design and construction of a simple underwater sensor network is described

to detect extreme temperature gradients (thermoclines), which are considered to be a

breeding groundfor certain marinemicroorganisms.

•

^UnderseaExplorations: UASNscanhelpdetectingunderwateroileldsorreservoirs,de- termineroutesforlayingunderseacables,andassistinexplorationforvaluableminerals.

•

^Disaster Prevention: WSNs that measure seismic activity from remote locations can provide tsunami warnings to coastalareas[21 ], orstudy theeectsof submarine earth-

quakes(seaquakes).

•

^{Forest Fire Detection: a SANET could be deployed to detect a forest re in its early}

stages[22 ]. Anumberofnodesneedtobepre-deployedinaforest. Eachnodecangather

dierenttypesofinformation fromsensors,suchastemperature,humidity,pressureand

position. Allsensingdataissentbymulti-hopcommunicationtothecontrolcentreviaa

numberofactuators(gatewaydevices)distributedthroughouttheforest. Theactuators

willbeconnectedtomobilenetworks(e.g.,UniversalMobileTelecommunicationsSystem

UMTS) andwill bepositionedsoasto reduce thenumberof hopsfromsource ofre

detection to the control centre. The actuators will also reduce network congestion in

large-scaledeploymentsbyextractingdatafromthenetworkatpredeterminedpoints. It

mayalso be possibleinthis scenariothatsomemobileforest patronunitsactasmobile

actuator, collecting environmental data asthey traverse throughtheforest. Assoonas

a re-relatedevent is detected,such assuddentemperature rise, thecontrol centre will

be alarmedimmediately. Operatorsinthecontrol centre can judge ifitisa falsealarm

byeitherusing thedatacollectedfromothersensors ordispatchingateam tocheckthe

situation locally. Thenboth reghters and helicopters can be sent to put out there

beforeit grows to a severeforest re.

1.2.4 Medical Applications

•

^Home Monitoring: home monitoring for chronic and elderly patients [23] allows long- term care andcan reduce thelengthof hospital stay.

•

^PatientMonitoring: sensornodesdeployedonthebodyofpatientsinhospitals[24 ]allow thecollection ofperiodicor continuousdata like temperature, bloodpressure, etc.

1.3 Motivations and Objectives

WSNs are similar to ad hoc networks in the sense that sensor networks borrow heavily on

the self-organizing and routing technologies developed by the ad-hoc research community.

However, a major design objective for sensornetworksisreducing thecost ofeach node. For

manyapplications, thedesiredcost fora wirelesslyenable deviceislessthan one dollar.

We, in this thesis, consider a set of sensors spread over a region to perform sensing op-

eration. Each of these sensors has a wireless transceiver that transmits and receives at a

single frequency, which is common to all these sensors. Over time, some of these sensors

generate/collect informationto besenttosome othersensor(s). Owingto thelimitedbattery

capacity of these sensors, a sensor may not be able to directly communicate with far away

nodes. In such scenarios, one of thepossibilitiesfor information transfer between two nodes

thatcannotcommunicate directlyisto useother sensornodesinthenetwork. Tobeprecise,

the source sensors transmits its information to one of thesensors which is within its trans-

missionrange. Theintermediatesensorthenusesthesameproceduresothattheinformation

nallyreachesits destination (a fusion center,i.e., a common sink 3

).

Asetcomprisingofordered pairofnodesconstitutea route thatisusedtoassistcommu-

nication between any two given pair of nodes (i.e., a sensor and a sink). This is a standard

problemof multihop routing inWSNs. Theproblem ofoptimal routing has been extensively

studiedinthecontext ofwirelinenetworkswhere usuallyashortestpath routingalgorithm is

used: each linkinthenetwork hasa weight associated withit and theobjective ofthe rout-

ingalgorithm is to nd a path that achieves the minimum weight between two given nodes.

Clearly, the outcome of such an algorithm depends on the assignment of weights associated

toeachlinkinthe network. Inwirelinecontext, therearemanywell-studied criteriato select

these weights for links such as the queueing delay. In WSNs, the optimality in the routing

algorithm isset to extend network lifetime (wherelifetime isdened asthetimespanned by

the network for some data aggregation till rst alive node gets disconnected due to energy

outage) inasinglesinknetwork. Innetworkswithmultiplesinks[25],theowissplitted and

sent to dierent basestations with theaim of extendingthe network lifetimeof these limited

battery WSNs. However, a complete understanding of the eect of routing on WSNs perfor-

mance and resource utilization (in particular, the stability of transmit buers and hence, the

end-to-enddelay and throughput) has not received muchattention.

Aftersensorsinthesensor/actuatorelddetect aphenomenon,theyeither transmit their

readingstothe resource-rich actuator nodeswhich canprocess allincoming dataandinitiate

appropriateactions,orroutedatabacktothesinkwhichissuesactioncommandstoactuators.

Weusetheformercaseinthisthesis. Theadvantageisthattheinformationsensedisconveyed

quickly from sensors to actuators, since they are close to each other. Moreover, since event

information is only transmitted locally through sensornodes, only sensors around the event

areaareinvolvedinthecommunicationprocesswhichresultsinenergyandbandwidthsavings

inSANETs.

Ifthemapping between asensornodeand one(or more)actuator 4

isgiven a priori,then

theproblemofndingoptimal minimumenergy routesto optimize networklifetimehasbeen

wellinvestigatedinthepast[25 , 26]for WSNs. But, thereis verylittle research contribution

toward nding optimal delay routes in SANETs. Further, in cases when there are multiple

actuators and mapping between the sensors and actuators is not given, the joint problem

of ndinga destination actuator and minimum end-to-end delay routes is a challenging and

interesting problem. Thisis becausethe end-to-enddelaysare topology dependent;actuator

selection basedonminimumhop routing alonecan not guarantee optimalend-to-end delays.

Further,inordertoprovideeectivesensingandacting tasks,ecientcoordinationmech-

anismsarerequired. Wewillmainlyfocusontwomostconstrainedcoordinationlevelsnamely:

sensor-actuator coordination, andactuator-actuator coordination. The most important char-

acteristic of sensor-actuator coordination is to provide low communication delay due to the

proximityofsensorsand actuators. However,since theroleofsinkdoesnot involvecollecting

thesensordataandcoordinatingtheactivitiesofthe nodes,sensorandactuator nodesshould

3

Byafusioncenteroracommonsink,wemeanalogicaldestinationfordata.Thiscanbelocatedanywhere

inoroutsidethenetworktopology.

4

actuators/basestationsareconsideredtohavesimilarsemanticsformodelingpurposes,i.e.,sinksfordata

generatedinthenetwork.

locallycoordinatewitheachothersoastoprovideecienttransmissionofsensorreadings. In

SANETs,forsensor-actuatorcoordinationthereisaneedtodevelopprotocolswhichareableto

providereal-timeserviceswithgiven delaybounds, accordingtoapplication constraintsanden-

sure an energy ecient communication among sensors andactuators. In SANETs, actuators

can communicate witheach otherinaddition to communicating withsensors. Sincethereare

few numberofactuator nodesandthepowercapacitiesof thesenodesarehigher thansensor

nodes, actuator-actuator communication is similar to the communication in wireless ad-hoc

networks. Actuator-actuatorcoordinationcanoccurinthecases wheretheactuatorreceiving

sensor data may not act on the event area due to small action range or insucient energy,

where one actuator may not be sucient to perform the required action, thus other nearby

actuators should be triggered, where multiple actuators receive the same event information

and there is an action threshold, hence these actuators should talk to each other so as to

decidewhichoneofthemperformstheactionandwheremultipleeventsoccursimultaneously.

Thus actuator should coordinate and communicate witheach other to performtask allocation

eciently andeectively.

We also considera SANETthatprolongs network lifetime byminimizing theenergy con-

sumption and, in parallel, takes care of delay-sensitivity of the sensed data. Therefore, in

cases, where there are multiple actuators and mapping between thesensors and actuators is

not given, the problem of nding an optimal actuator and extending network lifetime with

minimumend-to-enddelayconstraintsisaninterestingproblem. Thisproblemisrelevantfrom

boththeapplication's andwirelessnetworkingperspectives. Fromanapplicationrequirement

perspective,forsome real-timemultimediasensingapplications(e.g.,video surveillance),itis

necessarytohaveallthetracgeneratedfromasourcesensortoberoutedtothesameactua-

tor(albeitthatitmayfollowdierentroutes)sothatdecodingandprocessingcanbeproperly

completed. For multimediatrac suchasvideo, theinformation contained indierent pack-

ets from thesame source arehighly correlated and dependent. If thepackets generated by a

source are split and sent to dierent actuators, any of these receiving actuators may not be

able to decode the video packets properly. From awireless networking perspective, theactu-

ator chosenasasinkcouldhave asignicant impactontheend-to-enddelayswhichisahard

constraint [27] for sensor-actuator applications. As a result, there appears tobe a compelling

need to understand how to perform optimal routing to jointly achieve minimum end-to-end

delay routes andoptimize network lifetimein delay-energy constrained SANETs.

ApartfromSANETs,wealsoconsiderUASNswhicharedeployedtoperformcollaborative

underwater monitoringtasks. Thesensorsmustbe organized inanautonomous networkthat

self-congureaccordingtothevaryingcharacteristicsoftheoceanenvironment. Mostimpair-

mentsof the underwater acousticchannel are adequatelyaddressed at the physical layer, by

designingreceiversthatareabletodealwithhighbiterrorrates,fading,andtheinter-symbol

interference (ISI) caused by multipath. There were eorts at developing channel equalizers

and adaptive spatialprocessingtechniques sothatcoherent phasemodulationcan be usedto

achieve thedesired highspectral eciencies. These techniquesare computationally demand-

ing with manyparameter adjustments, and requirements thatare not especiallysuitable for

applications where autonomy,adaptability,and long-life batteryoperation arebeingcontem-

plated. Therefore, we analyze the factors that inuence acoustic communications in orderto

state the challenges posed by the underwater channels for underwater sensor networking.

1.4 Thesis Outline and Contributions

InChapter2,weconsideraWSNinwhichthesensornodesaresourcesofdelaysensitivetrac

thatneedstobetransferredinamulti-hopfashiontoacommonprocessingcenter. Weconsider

the following data sampling scheme: the sensor nodes have a sampling process independent

(layered architecture) of the transmission scheme as shown in Figure 1.3. This system is

like thepacketradio network (PRN)for which exact analysisis not available. We also show

that thestability condition proposedin the PRN literature is not accurate. First, a correct

stability condition for such a system is provided. Then, we proposed a cross-layered data

samplingscheme inwhich, thesensornodessamplenewdataonly whenithasaopportunity

(cross-layeredarchitecture) oftransmittingthedataasshowninFigure1.3. Itisalsoobserved

thatthis scheme gives a better performance in terms of delays and is moreover amenable to

analysis.

Figure1.3: ALayered andCross-LayeredArchitecture

To provide meaningful service such asdisaster and emergency surveillance, meetingreal-

time and energy constraintsand the stabilityat mediumaccess control (MAC) layerarethe

basic requirements of communication protocols in such networks. We also propose a cross-

layerarchitecturewithtwotransmitqueuesatMAClayer,i.e.,oneforitsowngenerateddata,

and theother for forwarding trac asshown in Figure1.4. We usea probabilistic queueing

discipline. Our rstmain resultconcerns the stability oftheforwardingqueuesat thenodes.

Itstatesthatwhetherornottheforwardingqueuescanbestabilized, byappropriatechoiceof

weighted fairqueueing (WFQ) weights, dependsonly on routing and channel access rates of

thesensors. Further,theweightsoftheWFQsplayaroleindeterminingthetradeo between

the power allocated for forwarding and the delay ofthe forwarding trac.

We then addressthe problemof optimal routing thataims at minimizing theend-to-end

delays. Since,weallowfortracsplittingatsourcenodes,weproposeanalgorithmthatseeks

theWardropequilibriuminsteadofasingleleastdelaypath. Wardropequilibriarstappeared

inthecontextoftransportationnetworks. Wardrop'srstprinciple states: Thejourneytimes

in all routes actually used are equal and less than those which would be experienced by a

single vehicle on any unused route. Each user non-cooperatively seeks to minimize his cost

of transportation. The trac ows thatsatisfythis principle areusually referred to as"user

equilibrium" (UE) ows, since each user chooses the route that is the best. Specically, a

user-optimized equilibriumisreachedwhennousermaylowerhistransportationcostthrough

Figure1.4: ASystem withTwo-Queues at MAC

unilateral action.

The distributed routing scheme is designed for a broad class of WSNs which converges

(in the Cesaro sense)to theset ofCesaro-Wardrop equilibria. Each linkis assigneda weight

and theobjectiveistoroutethroughminimumweight pathsusingiterative updatingscheme.

Convergence is established using standard results from the related literature and validated

by TinyOS simulation results. Our algorithm can adapt to changes in the network trac

and delays. The scheme is based on the multiple time-scale stochastic approximation algo-

rithms. The algorithm is simulated in TOSSIM and numerical results from the simulations

are provided.

InChapter3,weconsidera two-tier SANETand addresstheminimumdelayproblemfor

data aggregation. We analyzethe average end-to-end delayin thenetwork. The objective is

to minimize the total delay inthe network. We prove that this objective function is strictly

convex for the entire network. We then provide a distributed optimization framework to

achieve the required objective. The approach is based on distributed convex optimization

and deterministic distributed algorithm without feedback control. Only local knowledge is

used to update the algorithmic steps. Specically, we formulate the objective as a network

leveldelayminimizationfunctionwheretheconstraintsarethereception-capacity andservice-

rate probabilities. Using the Lagrangian dual composition method, we derive a distributed

primal-dual algorithm to minimize thedelay inthenetwork. We furtherdevelopa stochastic

delay-control primal-dual algorithm inthe presence of noisyconditions. We also present its

convergence and rateof convergence properties.

This chapter also investigates a delay-optimal actuator-selection problem for SANETs.

Each sensor must transmit its locally generated data to only one of the actuators. A poly-

nomial timealgorithm is proposedfor delay-optimal actuator-selection. We nally proposea

distributedmechanismforactuation control which covers alltherequirementsforan eective

actuation process.

In Chapter 4, we consider a three-tier SANET and present the design, implementation,

and performance evaluation of a novel low-energy, adaptive and distributed (LEAD) self-

organization framework. Thisframeworkprovidescoordination, routing,andMAClayerpro-

tocols fornetworkorganization andmanagement. Theframework isshowninFigure1.5. We

organizetheheterogeneous SANET into clusters where eachcluster ismanaged byan actua-

tor. To maximize the network lifetime and attain minimum end-to-enddelays,it is essential

tooptimally matcheach sensornode to anactuator and ndan optimal routingscheme. We

provide an actuator discovery protocol (ADP) that nds out a destination actuator for each

sensorinthenetwork basedon the outcome ofa cost function. Further,oncethedestination

actuatorsarexed,weprovideanenergy-optimal routingsolutionwiththeaimofmaximizing

networklifetime. Wethenproposeadelay-energy awareTDMAbasedMACprotocolincom-

pliancewiththeroutingalgorithm. Theactuator-selection,optimalrouting,andTDMAMAC

schemes together guarantees a near-optimal lifetime. The proposal is validated bymeans of

analysisandns-2 simulationresults.

Figure1.5: The LEADFramework

Delay and energy constraints have a signicant impact on the design and operation of

SANETs. Furthermore, preventing sensornodes frombeing inactive/isolated is very critical.

Theproblem of sensorinactivity/isolation arises fromthe pathloss and fading thatdegrades

the quality of the signals transmitted from actuators to sensors, especially in anisotropic

deploymentareas,e.g.,roughandhillyterrains. SensordatatransmissioninSANETs heavily

reliesontheschedulinginformationthateachsensornodereceivesfromitsassociatedactuator.

Therefore, ifthe signalcontainingscheduling information isreceived ata verylowpowerdue

to the impairments introduced by the wireless channel, the sensor node might be unable to

decodeit andconsequently it will remaininactive/isolated.

Sensors transmit their readings to the actuators. All actuators cooperate and jointly

transmit scheduling information to sensors with the useof beamforming. This results in an

important reductionofthe numberofinactivesensorscomparing tosingleactuator transmis-

sion for agivenlevelof transmit power. The reductionisdue to theresultingarray gain and

the exploitation ofmacro-diversity thatis provided bythe actuator cooperation. In order to

maximize network lifetimeandattain minimumend-to-enddelays,itisessential to optimally

match each sensor node to an actuator and nd an optimal routing solution. A distributed

solution for optimal actuator selection subjectto energy-delay constraints isalso provided.

InChapter5, we consideraUASNand rstanalyze amodulation schemeand associated

receiver algorithms. This receiver design take advantage of the time reversal 5

(TR) and

properties of spread spectrum sequences known as Gold sequences. Furthermore, they are

much less complex than receivers using adaptive equalizers. This technique improves the

signal-to-noise ratio (SNR) at the receiver and reduces the bit error rate (BER). We then

applied the phase conjugation to network communication. We show that this approach can

give almost zero BERfor atwo-hop communicationmode compared to thetraditionaldirect

communication. This linklayer information is usedat thenetwork layer to optimize routing

decisions. We showthese improvements by meansof analyticalanalysis andsimulations.

In Chapter 6, we present a general summary of the work achieved and the conclusions

concerning the results obtainedduring this thesis. Some perspectivesand openquestionsare

given for the continuation of this work in the area of cross-layer optimizations in wireless

sensor, sensor-actuator, andunderwater acousticsensornetworks.

5

Itisalsoknownasphaseconjugation(PC)inthefrequencydomain

Cross-Layer Routing in WSNs

Inthis Chapter, we considera WSN inwhich thesensor nodes aresources ofdelaysensitive

trac thatneedstobetransferredinamulti-hopfashiontoacommon processingcenter. We

rst consider the layered architecture. This system is like PRNs for which exact analysis is

not availableinthe literature. Wealso showthatthestabilityconditionproposedinthePRN

literature is not accurate. First, a correct stability condition for such a system is provided.

We thenproposea newdatasampling scheme: thesensornodessample newdataonly when

it hasanopportunity (cross-layered) oftransmitting thedata. It isobservedthatthis scheme

givesa better performanceinterms ofdelays andmoreoveris amenableto analysis.

We also propose a closed (cross-layered) architecture with two transmit queues at each

sensor

i

^{, i.e., one for its own generated data, and the other for forwarding trac. Our rst}

mainresultconcernsthestabilityoftheforwardingqueuesatthenodes. Itstatesthatwhether

ornottheforwardingqueuescanbestabilized(byappropriatechoiceofWFQweights)depends

only on routing and channel access rates of the sensors. Further, the weights of the WFQs

play a role in determining the tradeo between the power allocated for forwarding and the

delay oftheforwardingtrac.

We then addressthe problem ofoptimal routing thataims at minimizing theend-to-end

delays. Since we allow for trac splitting at source nodes, we propose an algorithm that

seeks the Wardrop equilibrium (i.e., the delays on the routes that are actually used by the

packets from a source areall minimumand equal) insteadof a single leastdelay path. Each

link is assigned a weight and the objective is to route through minimum weight paths using

iterative updating scheme. The algorithm is implemented in TinyOS Simulator (TOSSIM)

andnumerical resultsfrom thesimulation areprovided.

2.1 Introduction

WSNs are an emerging technology that has a wide range of potential applications including

environment monitoring, medical systems,robotic exploration, and smart spaces. WSNs are

becomingincreasinglyimportantinrecentyearsduetotheirabilitytodetectandconveyreal-

time, in-situ information for many civilian and military applications. Such networks consist

of largenumberofdistributedsensornodesthatorganizethemselvesintoa multihop wireless

network. Eachnodehasoneormoresensors,embeddedprocessors,andlow-powerradios,and

is normallybattery operated. Typically,these nodescoordinateto perform a common task.

We propose a closed (cross-layered) architecture for data sampling (application layer)

in a wireless sensor network. In this architecture, there is a strong coupling between the

sampling process andthechannel accessschemeasshowninFigure1.3. The objective inthe

closed architecture is to provide sucient and necessary conditions for the stability region

and reducing end-to-end delays. With mathematical analysis and simulations, we show that

the closed architecture outperforms the traditional layered scheme, both in terms of stable

operatingregion aswell astheend-to-end delays.

We also propose a closed architecture with two transmit queues for data sampling in a

wirelesssensornetwork. In thisarchitecture, weconsider anewdatasampling scheme: Node

i

^,

1 ≤ i ≤ N,

^{hastwo queuesassociated withit: one queue}

Q _i

^{contains thedatasampled by}

the sensornode itselfand the otherqueue

F _i

^{contains packetsthat node}

i

^{hasreceived from}

any of its neighbors and hasto be transmitted to another neighbor as shown in Figure 1.4.

In this architecture, there is coupling between the sampling process and the channel access

scheme. Theobjectiveintheclosedarchitectureistostudytheimpactofchannelaccessrates,

routing, and weights ofthe WFQson systemperformance.

We thenproposeanadaptive anddistributedrouting schemeforageneral classofWSNs.

TheobjectiveofourschemeistoachieveCesaroWardropequilibrium,anextensionoftheno-

tionofWardropequilibriathatrstappearedin[28 ]inthecontextoftransportationnetworks.

Wardrop's rst principle states: The journey times in all routes actually used areequal and

lessthan thosewhichwouldbeexperiencedbyasinglevehicleonanyunusedroute. Eachuser

non-cooperatively seeks to minimize hiscost of transportation. The trac ows that satisfy

this principleareusuallyreferredtoas"userequilibrium"(UE)ows,sinceeachuserchooses

the routethat isthe best. Specically, a user-optimizedequilibrium isreached when no user

may lower his transportation cost through unilateral action. The notion is dened in (2.1)

later in this chapter. Our algorithm is actually an adaptation of the algorithm proposed in

[29 ]tothecaseofWSNs. Inthealgorithm of[29],eachsourceusesatwotime-scalestochastic

approximation algorithm. Dierences inthe two algorithmsare:

1. In WSNs that we consider, each node has an attribute associated with it namely the

channel access rate. The delay on a route depends on the attributes of the nodes on

the route. However, in orderto maintain some longterm data transferrate, each node

needs to adaptits attributeto routing.

2. The dierence intime scalesthatwe usefor various learning/adaptation schemes helps

us prove convergence of ouralgorithm [C-4](sucha proof isnot present in[29 ]).

In this thesis, we consider a static wireless sensor network with

n

^{sensor nodes. Given is}

an

n × n

neighborhood relation matrix

N

^{that indicates the node pairs for which direct}

communication ispossible. We willassume that

N

^{isa symmetric1 matrix, i.e.,ifnode}

i

^can

transmitto node

j

^,then

j

^{canalso transmit tonode}

i

^{. For suchnode pairs,the}

(i, j) ^th

^entry

of the matrix

N

^{isunity,i.e.,}

N _i,j = 1

^{if node}

i

^and

j

^can communicate witheach other; we willset

N _i,j = 0

^ifnodes

i

^and

j

^{can not} communicate. For anynode

i

^{,we dene}

N _i = { j : N _i,j = 1 } ,

Whichis theset ofneighboring nodesof node

i

^{. Similarly, thetwo hop neighbors ofnode}

i

^{are dened as}

1

Theassumptionofsymmetryistoonlydrivetheanalysis. Weconsiderassymmetriclinksforconducting

simulations.

S _i = { k / ∈ N _i ∪ { i } : N _k,j = 1 f or some j ∈ N _i }

Notethat

S i

^{doesnot include anyof the rst-hopneighborsof node}

i

^.

Eachsensornodeis assumedto be sampling(or, sensing) itsenvironment at a predened

rate; we let

λ i

^{denote this sampling rate for node}

i

^{. The units of}

λ i

^{will be packets per}

second, assuming same packet size for all the nodes in the network. In this work, we will

assume that the readings of each of these sensor nodes are statistically independent of each

other so that distributed compression techniques are not employed (see [30] for an example

wheretheauthorsexploitthecorrelationamongreadingsofdierentsensorstousedistributed

Slepian-Wolf Coding[31] to reducethe overall transmissionrateof thenetwork).

Eachsensornodewantstousethesensornetworktoforwarditssampleddatatoacommon

fusion center (assumed to be a part of the network 2

). Thus, each sensor node acts as a

forwarder ofdatafrom othersensor nodesinthenetwork. We willassume thatthebuering

capacity of each node is innite 3

,sothat there isno data loss inthenetwork. We will allow

for thepossibility thata sensor node discriminates between its own packets and thepackets

to be forwarded(thus allowing forthemodelof[32]whichconsidersanAdHocnetwork. The

nodes in this network probabalistically schedule their transmissions to discriminate between

the fowarding trac and the one generated bynode itself).

Welet

φ

^denotethe

n × n

^{routingmatrix. The}

(i, j) ^th

^{elementofthismatrix,denoted}

φ _i,j

^,

takes valuein the interval

[0, 1].

^{This means a} probabilistic ow splittingas inthe model of [33 ],i.e., afraction

φ i,j

^ofthetrac transmitted fromnode

i

^{isforwardedbynode}

j

^asshown

in Figrue 2.1. Clearly, we need that

φ

^{is a stochastic matrix, i.e., its row elements sum to}

unity. Also notethat

φ _i,j > 0

^{ispossibleonly if}

N _i,j = 1

^.

Figure2.1: FlowSplitting

We assume that the system operates in discrete time, so that the time is divided into

2

Conceptually,wecanassumethat thisfusioncenterisalsoasensornode,whichhas

0

^{samplingrate. A}

negativesamplingratewouldmeanpushingdatafromthenetworktowardsthefusioncenter.

3

We assumeinnitebuersizeonlyto keepthe analysissimple. Later,weconsiderxedbuersizesand

lookatvarioustypesofdatalosses.